csm Library

Note

The library must be initialized with csm_init first, before it can be used.

The csm library provides a unified interface for the following types of functions:

  • linear functions y= a + b*x

  • functions of the form y= f(x) defined by a lookup table

  • functions of the form z= f(x, y) defined by a lookup table

All these functions are represented with a single abstract data type, csm_function.

A csm_function object must always be created with csm_new_function.

The csm library provides functions to manage csm_function objects and to do calculations. All calculation functions cache the last value and are optimized for speed.

Linear functions

This is the most simple type of function of the type y= a + b*x. It is defined with csm_def_linear.

One dimensional functions with lookup table

This implements a function y= f(x) with a lookup table. The lookup table can be initialized by a file or from two arrays of type double. If the function f(x) is monotonic, you can also calculate the inverse function x= g(y).

The one dimensional lookup table

These are requirements/properties for a one dimensional table:

  • A line must not be longer than 127 bytes.

  • A line that is not empty or a comment must have exactly two numbers

  • A line with numbers may start or end with an arbitrary number of white spaces

  • Lines do not need to be ordered by x or y.

  • One value of x must never be mapped to two different values of y

Here is a short example of a table:

5.5000   641.0683
5.6000   657.0219
5.7000   673.1084
5.8000   689.3115
5.9000   705.6146
6.0000   722.0005
6.1000   738.4555
6.2000   754.9796
6.3000   771.5765
6.4000   788.2499

CSM uses linear interpolation to compute the function y= f(x) for all possible values of x (see csm_x). If the function f(x) is monotonic, you can also compute the inverse function x= g(y) (see csm_y).

For x values smaller than the first value in the table or larger than the last value in the table, the function is linear extrapolated.

In the example above, we have:

f(6.3)  = 771.5765
f(6.35) = 779.9132
f(6.4)  = 788.2499
f(6.45) = 796.5866

Two dimensional functions with lookup table

This implements a function z= f(x, y) with a lookup table. The lookup table is always initialized by reading a file.

The two dimensional lookup table

A two-dimensional table specifies a function with two parameters. The first line and the first column specify the XY-grid. The first line must have exactly one element less than the remaining lines; it specifies the Y coordinates of the grid. The first column (i.e. the first elements of the remaining lines) specify the X coordinates of the grid, while the remaining elements specify the value (Z coordinate) at the corresponding point in the grid (see csm_z).

These are requirements for a two dimensional table:

  • A line must not be longer than 1023 bytes.

  • The first line must contain all y values

  • All following lines must contain a single x and all z values for z= f(x, y).

  • Lines and columns do not need to be ordered by x or y.

Here is a short example of a table:

         -6.0000  -3.0000  0.0000   3.0000   6.0000
15.4000  -0.1990  -0.1150  0.0000   0.1240   0.2580
15.9000  -0.1990  -0.1150  0.0000   0.1240   0.2580
16.6000  -0.2050  -0.1230  0.0000   0.0830   0.2190

CSM uses two dimensional linear interpolation to compute the function z= f(x, y) for all possible values of x and y (see csm_z).

In the example above, we have:

f(15.9, -6) = -0.1990
f(15.9, -3) = -0.1150
f(16.6, -6) = -0.2050
f(16.6, -5) = -0.1777
f(16.6, -3) = -0.1230

Full API documentation

Here is the full API documentation:

Doxygen generated API docs,